Studies of efficient and stable organic solar cells based on aluminum-doped zine oxide transparent electrode

Transcription

1 Hong Kong Baptist University HKBU Institutional Repository Open Access Theses and Dissertations Electronic Theses and Dissertations Studies of efficient and stable organic solar cells based on aluminum-doped zine oxide transparent electrode Hanxiao Liu Hong Kong Baptist University Follow this and additional works at: Recommended Citation Liu, Hanxiao, "Studies of efficient and stable organic solar cells based on aluminum-doped zine oxide transparent electrode" (2014). Open Access Theses and Dissertations This Thesis is brought to you for free and open access by the Electronic Theses and Dissertations at HKBU Institutional Repository. It has been accepted for inclusion in Open Access Theses and Dissertations by an authorized administrator of HKBU Institutional Repository. For more information, please contact

2 Studies of Efficient and Stable Organic Solar Cells Based on Aluminum-doped Zinc Oxide Transparent Electrode LIU Hanxiao A Thesis submitted in partial fulfillment of the requirements for the degree of Master of Philosophy Principal Supervisor: Prof. ZHU Fu Rong HONG KONG BAPTIST UNIVERSITY August 2014

3

4 Declaration I hereby declare that this thesis represents my own work which has been done after registration for the degree of M.Phil. at Hong Kong Baptist University, and has not been previously included in a thesis or dissertation submitted to this or any other institution for a degree, diploma or other qualifications. Signature: Date: August, 2014 i

5 ii

6 Abstract Organic solar cells (OSCs) have attracted significant attention due to their potential of large area solution fabrication capability at low-cost. For bulk heterojunction (BHJ) OSCs, a thin film of transparent conducting indium tin oxide (ITO), coated on glass or flexible plastic substrate, is widely used as a front electrode. However, indium is not abundant on Earth. Its price has increased continuously over the past 10 years and will likely become an obstacle for the commercialization of OSCs at low cost. Aluminum-doped zinc oxide (AZO) is a promising ITO alternative due to its advantages of high electric conductivity, optical transparency, non-toxicity and low cost. However, reports on OSCs using AZO electrode are quite limited, due to the relatively lower power conversion efficiency (PCE) of AZO-based OCSs as compared to that of ITO-based OCSs. This work focused on studies of high performance AZO-based OSCs through AZO surface modification, absorption enhancement and process optimization. The optical and electronic properties of AZO film including transmittance, sheet resistance, surface morphology and surface work function were characterized. AZO-based OSCs with conventional and inverted structures were fabricated. It was found that AZO-based OSCs with inverted structure demonstrated superior performance than the ones with conventional structure. The inverted structure avoids the use of acidic PEDOT:PSS hole transporting layer, allows the improving of the absorbance of the OSCs and therefore its efficiency. An AZO front transparent cathode was used for application in high performance inverted BHJ OSCs. The photoactive layer consisted a blend of poly[[4,8-bis[(2- ethylhexyl)oxy] benzo [1,2-b:4,5-b'] dithiophene-2,6- diyl][3-fluoro- 2-[(2-ethylhexyl) carbonyl]thieno[3,4-b]thiophenediyl]](ptb7):3'h-cyclopropa[8,25][5,6]fullerene- C70- D5h(6)-3'-butanoicacid, 3'-phenyl-, methyl ester (PC 70 BM). A structurally identical control OSC having an ITO front cathode was also fabricated for comparison studies. The structure of OSCs was optimized to achieving absorption enhancement in iii

7 the active layer. AZO and ITO were modified with a 10 nm thick solution-processed ZnO interlayer to facilitate the efficient electron extraction. The results revealed that bilayer AZO/ZnO and the ITO/ZnO cathodes possess similar electron extraction property. AZO layer has a transparency cutoff at wavelength < 380 nm, results in a slight decrease in the short-circuit current density (J SC ). However, the decrease in J SC is very small because the main energy of solar irradiation falls in the spectrum with wavelength > 380 nm. It shows that AZO-based OSCs have a promising PCE of 6.15%, which is slightly lower than that of a control ITO-based OSC (6.57%). AZO-based OSCs, however, demonstrate an obvious enhancement in the stability under an ultraviolet (UV)-assisted acceleration aging test. The significant enhancement in the stability of AZO-based OSCs arises from the tailored absorption of AZO electrode in wavelength < 380 nm, which serves as a UV filter to inhibit an inevitable degradation process in ITO-based OSCs due to the UV irradiation. In order to further investigate the degradation mechanism of OSCs under UV exposure, the change in charge collection characteristics of the OSCs made with ITO/ZnO and AZO/ZnO front cathode before and after UV exposure was examined. It was found that there was an obvious decrease in the charge extraction efficiency of ITO-based OSCs after UV exposure, while there was no observable change in the charge extraction efficiency of OSCs made with AZO/ZnO cathode under the same acceleration aging test. This work demonstrates that AZO is a suitable ITO alternative for application in OSCs, offering an improved device stability, comparable PCE and cell fabrication processes with an attractive commercial potential. iv

8 Acknowledgement I would like to express my deepest and sincere gratitude to my thesis advisor, Professor Furong Zhu, for his enlightening guidance and instruction in the development and completion of this study. Professor Zhu stimulated my interest in the field of organic electronics, providing not only prudent and pointed guidance but also gave me warm and sincere support. I am very appreciative of my group members, Mr. Zhenghui Wu, Dr. Bo Wu, Mr. Wing Hong Choi, Dr. Hoi Lam Tam and Mr. Qingyi Yang, for the many fruitful discussions and assistances in experiments. Their support and encouragement played a great deal towards the motivation of my study. My gratitude also goes to visiting scholars, Dr. Qunliang Song and Dr. Jianqiao Hu who have helped me with the publication of my first paper, and Dr. Wenyu Ji for the useful discussions. I would also like to express my appreciation to the scientific staff at Department of Physics, HKBU, Mr. Edward Chan, Ms. Winnie Wu and Mr. Benson Leung for their continuous support and devoted time in ensuring the quality and maintenance of lab equipment. A special gratitude and love goes to my family for their unfailing support and encouragement during my study. LIU Hanxiao Hong Kong Junly, 2014 v

9 List of Abbreviations OSC OPV BHJ TCO ITO AZO HTL ETL EEL PTB7 Organic solar cell Organic photovoltaic Bulk heterojunction Transparent conducting oxide Indium tin oxide Aluminum-doped zinc oxide Hole transport layer Electron transport layer Electron extraction layer Poly[[4,8-bis[(2-ethylhexyl)oxy]benzo[1,2-b:4,5-b']dithiophene-2,6-diyl][3-fluoro-2-[(2-ethylhexyl)carbonyl]thieno[3,4-b]thiophen ediyl]] P3HT ICBA PC 70 BM Poly(3-hexylthiophene-2,5-diyl) Indene-C 60 bisadduct 3'H-Cyclopropa[8,25][5,6]fullerene-C 70 -D5h(6)-3'-butanoicacid, 3'-phenyl-, methyl ester PC 60 BM PEDOT:PSS CB DIO C 60 derivative, [6,6]-phenyl-C 61 -butyric acid methyl ester Poly(3,4-ethylenedioxythiophene) poly(styrenesulfonate) Chlorobenzene 1, 8-Diiodooctane vi

10 HOMO LUMO J V FF Highest occupied molecular orbital Lowest unoccupied molecular orbital Current density voltage characteristics Fill factor V OC Open-circuit voltage J SC Short-circuit current density J ph Photocurrent density V eff Effective voltage PCE IPCE EQE VASE n( ) k( ) AFM UPS Power conversion efficiency Incident photon-to-electron conversion efficiency External quantum efficiency Variable angle spectroscopic ellipsometry Wavelength dependent refractive index Wavelength dependent extinction coefficient Atomic force microscopy Ultraviolet photoelectron spectroscopy vii

11 Table of Contents: Pages Declaration... i Abstract... iii Acknowledgement... v List of Abbreviations... vii Chapter 1: Introduction Background and motivation Organic photovoltaics Brief overview of the development of solar cells Prospects and challenges Thesis objectives... 9 Chapter 2: Overview of organic solar cells Basics of organic solar cells Device architectures and materials Characterization of OSCs Equivalent circuitry diagram Current density Voltage characteristics IPCE measurement Chapter 3: Experimental Materials process and characterization Optical properties of functional layers Sheet resistance and resistivity of AZO Chapter 4: Optical design of organic solar cells Optical admittance analysis Device design and optical modeling Chapter 5: Organic solar cells with AZO cathode Recent progresses in AZO-based OSCs Modification of ZnO with solution-processed ZnO Device characterization and analysis Stability of AZO-based organic solar cells Charge extraction efficiency of AZO-based OSCs Correlation between the change in charge collection 55 1

12 characteristics and the stability of the OSCs... Chapter 6: Summary and future work Summary Future work References Curriculum Vitae

13 List of Figures Fig. 1.1 Rates of world energy usage... 2 Fig. 1.2 Progress of solar cells... 4 Fig. 1.3 Cross-sectional views showing the structures of (a) single layer OSC, (b) bi-layer OSC and (c) bulk-heterojunction OSC... 6 Fig. 2.1 Charge generation and transfer processes in an OSC Fig. 2.2 Device structures of conventional (a), and inverted (b) OSCs Fig. 2.3 Chemical structures of PTB7, PC 70 BM, CB and DIO Fig. 2.4 Equivalent circuitry diagram of a solar cell Fig. 2.5 J V characteristics of an OSC Fig. 2.6 Definition of air mass Fig. 2.7 Spectra of AM 1.5 Global and AM 1.5 Direct Fig. 2.8 Schematic diagram of J V measurement system setup Fig. 2.9 Schematic diagram of IPCE measurement system setup Fig. 3.1 (a) schematic diagram of a UV-Vis spectroscopy and (b) a photo picture of the UV-Vis spectroscopy used in this work Fig. 3.2 Absorption spectra of PTB7:PC 70 BM films with different weight ratios of PTB7:PC 70 BM in the blend Fig. 3.3 Configuration of a VASE Fig. 3.4 Wavelength-dependent n( ) and k( ) measured for functional materials e.g. ITO, blend layer of PTB7:PC 70 BM, ZnO and MoO x used in this work Fig. 3.5 AZO sheet resistance & resistivity vs. layer thickness Fig. 4.1 Cross sectional view of (a) conventional and (b) inverted OSCs based on PTB7:PC 70 BM blend system Fig.4.2 Calculated integrated absorptance of active layer as a function of active layer thickness for conventional and inverted OSCs Fig. 4.3 Electric field distribution in PTB7:PC 70 BM based inverted OSCs with active layer thickness of (a) 60 nm and (b) 110 nm Fig. 5.1 The pattern of ITO/AZO on glass substrate Fig. 5.2 The design of a completed organic solar cell Fig. 5.3 Cross-sectional view of an inverted OSC made with a front 3

14 transparent cathode Fig. 5.4 J V characteristics of inverted OSCs made with different cathodes of AZO (850 nm)/zno (10 nm), ITO (170 nm)/zno (10 nm) and bare AZO (850 nm), measured under simulated AM1.5G, 100 mw/cm 2 illumination Fig. 5.5 Transparency measured for an AZO/glass and an ITO/glass over the wavelength range from 320 nm to850 nm, spectrum of AM1.5G solar irradiance is also plotted Fig. 5.6 IPCE of inverted OSCs made with an AZO/ZnO (10 nm) transparent cathode and a control OSC having an ITO/ZnO (10 nm) cathode Fig. 5.7 Schematic energy diagram of an AZO-based inverted OSC Fig. 5.8 AFM images measured for (a) glass/azo and (b) glass/azo/zno Fig. 5.9 Recorded PCE, FF, J SC and V OC of inverted OSCs made with ITO and AZO front cathodes versus the UV exposure time Fig Illustration of compensation voltage V Fig J ph V eff characteristics of inverted OSCs made with different cathodes of AZO/ZnO and ITO/ZnO Fig J ph V eff characteristics of inverted OSCs made with different cathodes of AZO/ZnO and ITO/ZnO Fig J ph as a function of light intensity for OSCs made with cathodes of (a) AZO/ZnO and (b) ITO/ZnO. The slopes of fitted lines and corresponding V eff are also shown Fig Recorded J ph V eff characteristics of AZO-based and ITO-based OSCs under different light intensity (a) before UV exposure and (b) after 10mins UV exposure Fig J ph as a function of light intensity for OSCs made with cathode of (a) AZO/ZnO before UV exposure, (b) AZO/ZnO after 10 mins UV exposure, (c) ITO/ZnO before UV exposure and (d) ITO/ZnO after 10 mins UV exposure. The slopes of fitted lines and corresponding V eff are also shown

15 List of Tables Table 1 Summary of results on OSCs using AZO as electrode Table 2 Summary of the performance of OSCs made with different transparent cathodes of AZO/ZnO, ITO/ZnO and AZO

16 Chapter 1: Introduction 1.1 Background and motivation We are in an age of remarkable advancements in industrialization and technology, where the average income exhibits unprecedented sustained growth, leading to a rapid increase in population. This influences the living standards of the masses of ordinary people which in turn cause an increase in global energy consumption. What we see today is a high dependence on fossil fuels such as coal, natural gas and petroleum to meet basic human needs. However it is becoming increasingly obvious that there are problems on this heavy reliance which we need to address. Not only is there a limited amount of fossil fuel, there is environmental damage in getting the fossil fuels out of the ground and when they are burnt as a fuel source. The burning of these fuels produces air pollution. It also releases CO 2 into the atmosphere which can cause global warming. The concern of environmental pollution and energy shortage are the driving force for the searching for alternative energy sources. There are various alternative energy sources, including nuclear energy, hydroelectricity and renewable energies such as wind energy, geothermal energy, biofuel and solar energy. Although nuclear energy is considered as a promising alternative, there remains the concern for nuclear pollution and nuclear waste disposal. Hydroelectricity has low power cost and no CO 2 emission, but it can be detrimental to the ecosystem and posts a failure risk. 1

17 Fig. 1.1 Rates of world energy usage. While we eagerly search for renewable and clean energy sources, the most abundant one is just around us. Solar energy is the largest energy source on Earth. Approximately the total solar energy of J per year is absorbed by Earth's atmosphere, oceans and land masses [1]. At the peak hour, the sun can deliver more energy in one hour than the world uses in one year [2].If 0.1% of the earth surface is covered with the panels of solar cells with power conversion efficiency (PCE) of 10%, the electricity power generated can satisfy the present energy needs[3]. Although it is not realistic to cover such ground with solar cells at present, this prediction gives a good idea of the abundance of solar energy. Solar energy is essentially electromagnetic waves. It is clean, safe and practically limitless. There are many ways to utilize this energy. One way is to harvest it as thermal energy to heat buildings or waters for domestic use. The second is to convert it into electricity using photovoltaic effect. 2

18 1.2 Organic photovoltaics Brief overview of the development of solar cells Photovoltaic effect was first discovered in 1839 by Edmund Becquerel who observed a photo induced current in a device consisting of two electrodes [4]. While the photovoltaic effect was studied subsequently throughout the century, the first modern silicon solar cell, based on a p-n junction, was first demonstrated by Chapin et. al. at Bell Labs in 1954 [5]. The maximum efficiency of single p-n junction solar cells is approximately 31% due to the Shockley-Queisser limit [6]. To date, considerable progress has been made in PCE of difference types of solar cells, e.g., crystalline silicon cells (25.0±0.5%), III-V (GaAs) semiconductor solar cells (26.1±0.8%), thin film amorphous silicon solar cells (9.5±0.3%) [7], organic solar cells (~10%) [8] and more recently perovskite solar cells (>15%) [9]. 3

19 Fig. 1.2 Progress of solar cells. (On the right bottom side is the record PCE reported for OSCs). Despite the rapid development of photovoltaic in the past half century, solar energy still remains expensive compared to other energy sources. The high cost originated from the expensive materials (Si), high cost fabrication processes (high vacuum) and capital investment. Based on data from International Energy Agency, the total electricity generated globally in 2008 was 20,261TWh, the amount generated by solar cells was 12TWh, counting only 0.06%. The emerging OSC technology has the potential to overcome the cost drawback of inorganic solar cells. OSCs are based on organic small molecules or conjugated polymers, which are low-cost and suitable for massively production. They are compatible to vacuum-free solution-processed fabrication technique with a potential of large-area-roll-to-roll production at low cost. It also has many other attractive 4

20 advantages such as mechanical flexibility, light weight and semi-transparency. This opens up new photovoltaic designs and products for application in ultra-thin flexible portable solar cells or energy generating windows. Another advantage of OSCs is its tunable absorption spectrum. By changing the chemical structure of organic photoactive materials, the absorption spectrum of OSCs can be tuned for more versatile applications. Organic photovoltaic started with the discovery of photovoltaic effect of chlorophyll in plant by C. W. Tang in 1975 [10]. However, in the following few years, the PCE of OSCs was very low due to the limited charge separation at the organic/metal interface. The introduction of bi-layer heterojunction structure in 1986 also by C. W. Tang helped to make a big step forward [11]. In this bi-layer structure, n-type (acceptor) and p-type (donor) layers are sandwiched between two electrodes with different work functions. The LUMO level of the acceptor is slightly lower than that of the donor, promoting exciton dissociation at the donor/acceptor interface with much higher dissociation efficiency, resulting in PCE of 1%. But this is still quite low compared to inorganic solar cells, limited by the short exciton diffusion length in most organic photoactive materials. If the photo-generated excitons cannot reach the donor/acceptor interface where the dissociation of excitons takes place, they will loss to the different recombination processes. Light absorption in OSCs is limited due to a mismatch between the effective thickness of the functional layer and the poor carrier mobility in organic semiconductors, therefore the PCE. Further improvement was made with the introduction of bulk-heterojunction structure in the mid 1990s, in which the p-type 5

21 polymer and the n-type fluorine are mixed to form a blend layer. This configuration makes the length scale of the blend to be around a few tens of nanometers, in the same order of magnitude as the diffusion length of excitons. Therefore, the photo-generated excitons with the bulk heterojunction are able to diffuse to the donor/acceptor for efficient dissociation before lost to the recombination processes. This greatly increases the chance of charge separation, allowing employing relatively a thicker active layer in OSCs for improved light absorption. Fig. 1.3 Cross-sectional views showing the structures of (a) single layer OSC, (b) bi-layer OSC and (c) bulk-heterojunction OSC. Significant progresses in OSCs with a PCE of > 9% have been reported [11]. In 2010, a polymer material named PTB7 designed by Y.Y. Liang et.al. demonstrated a 7.4% PCE (the highest at that time) [12]. Until now, PTB7 is one of the most widely used p-type polymers in OSCs, with a broad absorption spectrum for achieving high PCE. In 2012, an inverted OSC utilizing a PFN electron transport layer achieved 9.2% with PTB7:PC 70 BM blend system [13], close to the theoretical 10% threshold for OSCs to enter commercial market Prospects and challenges Although encouraging progresses have been made, considerable enhancement in overall performance of OSCs, including efficiency, durability and cost 6

22 competitiveness is needed if this technology is to become a viable option for sustainable energy [14,15]. Although organic photovoltaic technology offers an attractive option for achieving alternative clean energy sources, the use of indium-tin-oxide (ITO) transparent electrode is not a long-term solution for large area OSCs at low cost [16,17]. This is because indium is not abundant on Earth. The limited indium resource and large demand for ITO in flat panel display industry, e.g., liquid crystal displays and more recently flat panel displays based on organic light emitting diode (OLED), has led to a doubling in demand for the ITO materials. There remains a need for the development of alternatives to ITO for application in OSCs. A variety of ITO replacements have been investigated. For example, Metal grid using roll-to-roll technique has been demonstrated by F.C. Krebs [18]. But the performance of OSCs with these ITO-free contacts is somewhat inferior to those with ITO because of the poor light transmission. Carbon-based nanostructures such as carbon nanotubes [19,20] and graphene [21,22] are also investigated as ITO replacements in OSCs. But the performance of devices is also low compared to ITO based ones due to the trade-off between transmission and conductivity of these replacement materials. Other ITO replacements include high conductivity Poly(3,4-ethylenedioxylenethiophene): poly (styrenesulphonic acid) (PEDOT: PSS) [23], and impurity-doped-zno [24]. Among them, the doped-zno has received a lot of attention due to its advantages of abundance in nature and nontoxicity. The properties of ZnO films can be easily modified by incorporating a variety of extrinsic dopants, for example, B, Al, In, Ga, Si, Sn, F and Cl. Among these impurity-doped ZnO films, aluminum-doped zinc oxide 7

23 (AZO) films exhibit both high visible transmittance and low resistivity that are comparable with that of the ITO films. Thin films of doped-zno are practically attractive for use as a transparent conductive coating, the doped-zno films can be readily produced for large scale coating at low cost, allow tailoring of the ultraviolet (UV) absorption and good thermal stability. AZO is a promising alternative to ITO for application in OLEDs and OSCs due to its unique electrical and optical characteristics [25,26]. Apart from cost, stability is also an important issue for OSCs. It is known that organic solar cells undergo continuous degradation when exposed to moisture, oxygen and ultraviolet (UV) light. Traps will be introduced in the bulk, creating sites for charge recombination. Interfacial defects will reduce charge extraction efficiency, leading to charge accumulation and space-charge-limit-current (SCLC). It is essential to find ways to prevent or slow these degradations in OSCs for OSCs to become commercially attractive. 8

24 1.3 Thesis objectives Research on OSCs has been largely focused on improving the energy conversion efficiency, which is the prerequisite for using OSCs in energy generation. However, if OSC is to become a commercially viable long-term solution for our society s energy need, it not only needs to have high energy conversion efficiency, but also long operation lifetime and low production cost. As mentioned above, the use of ITO contributes a big proportion to the total cost of OSCs. This will become an obstacle for the commercialization of low-cost OSCs. AZO offers a lot of attractive characteristics and can be used as an ITO alternative. However, study on AZO-based OSCs is quite limited, due to the overall inferior performance of AZO-based OSCs as compared to that of the ITO-based OSCs. The aim of this research was to carry out a systematic study to investigate AZO as an ITO replacement for application in high performance OSCs. The research had the following goals: 1. To optimize the performance of conventional and inverted OSCs on ITO substrates using theoretical simulation and experimental optimization. 2. To develop high performance AZO-based OSCs based on the optimized structure and fabrication processed derived from ITO-based OSCs. 3. To analyze the advantages and limitations of AZO-based OSCs, and optimize the performance of AZO-based OSCs. 4. To study the device stability of ITO-based and AZO-based OSCs. 9

25 Chapter 2: Overview of organic solar cells 2.1 Basics of organic solar cells The basic working principle of energy generation in an OSC is different from its inorganic counterpart. In an OSC, due to the low dielectric constant in organic semiconductor materials as compared to conventional silicon, the photo-excited excitons have much higher binding energy compared to that in Si. The binding energy of photo-generated excitons in organic materials is in a range from 200 to 500 mev, about 10 times larger than the ones in conventional inorganic semiconductors, e.g., Si, where photo-excitations typically lead to free carriers directly at room temperature. The photo-excited excitons diffuse to the donor/acceptor interface and are separated under the potential difference between the LUMO levels of the donor and acceptor materials. The dissociated electrons and holes are then drifted to the cathode and anode under the built-in electric field, created by the difference in the work function between the two electrodes. The processes of charge generation in an OSC are shown in Fig. 2.1 below. Fig. 2.1 Charge generation and transfer processes in an OSC. 10

26 The energy conversion processes in an OSC can be summarized into 4 steps: Step (a): Light absorption and exciton generation: Organic materials absorb incident light and electrons are excited from HOMO level to LUMO level, creating excitons. Step (b): Exciton diffusion: Excitons diffuse to the interface of donor (p-type polymer) and acceptor (n-type fullerene and its derivatives). Step (c): Charge separation Dissociation of excitons takes place at donor/acceptor interface to form charge transfer states, generating electrons and holes. Step (d): Charge transportation: Electrons travel to cathode while holes to anode under the built-in potential. 2.2 Device architectures and materials There are two common structures for OSCs, namely, the conventional and the inverted structures. In both types of architectures, the active layer is sandwiched between two electrodes. For work function alignment and charge transport purpose, hole transporting layer (HTL) and electron transporting layer (ETL) are often inserted between the active layer/anode and active layer/cathode interface respectively. The front electrode needs to be transparent to sunlight, while the back electrode is normally reflective to increase optical path length and therefore enhance light absorption. Both electrodes are required to have good electrical conductivity. In 11

27 conventional structure OSCs, light incidents through the anode side of the cells. The most commonly used anode material is ITO. Oxygen plasma treatment is often used to increase the work function of ITO from 4.5 ev to ~5.0 ev. Between the anode and active layer, a HTL is used to block of electrons and transport holes. Currently PEDOT:PSS is the most popular HTL material for conventional structured OSCs. PEDOT:PSS can be solution processed in ambient condition, which is a simple fabrication process to produce a smooth electrode surface, making it easier for the subsequent coating of active layer. At the back, an ETL is used to block holes and extract electrons. The commonly used ETL materials include LiF, ZnO and TiO x. Finally, low work function metals, such as Li, Ca, Ag and Al, are used as a cathode. For the inverted structure, light incidents through the front transparent cathode side of the cells. ITO can also be used as the cathode. But because of its high work (~4.5 ev), an ETL is required to modify it for cathode application. For the HTL, MoO x is often used before evaporating Ag or Au as anode. Fig. 2.2 Device structures of conventional (a), and inverted (b) OSCs. For bulk-heterojunction OSCs, the blend layer is typically made up of a mixture of p-type polymer donor and n-type fullerene acceptor. The blend layer will absorb incoming photons to generate excitons. Only photons with energy larger than the 12

28 optical band gap of the functional photoactive materials can be absorbed. For example, a small band gap donor material will have a broader optical absorption spectrum and a favorable exciton generation rate. The voltage output of the OSCs is closely related to the energy difference between the HOMO level of donor and LUMO level of acceptor material. If the band gap of the donor material decreases, the output voltage of the OSCs will also decrease. Since the overall PCE of OSCs is determined by both the photocurrent and voltage, this means a balance needs to be achieved between the current and voltage output to maximize the PCE. It is possible to raise the LUMO level of acceptor material to increase the voltage without sacrificing light absorption. For example, in a previous work reported by Li et.al. [27], in the donor system of Poly(3-hexylthiophene-2,5-diyl) (P3HT), by changing the acceptor material from C 60 derivative, [6,6]-phenyl-C 61 -butyric acid methyl ester (PC 60 BM) which has a LUMO energy of ev to indene-c 60 bisadduct (ICBA) with a higher LUMO energy of ev, the voltage output increased without affecting the current generation. This leads to an increase in PCE from 3.73% to 6.48%. However, for exciton dissociation to occur at the D/A interface, an energy potential difference is required. It has been reported that a minimum energy difference of 0.3 ev between the LUMO levels of donor and acceptor is needed to separate excitons at the D/A interface. Apart from energy levels, morphology of the functional layers also plays an important role in determining the cell performance. After excitons are dissociated into free electrons and holes, a percolation pathway is needed to transport the free carriers to 13

29 the respective electrode. The morphology or phase separation of the active layer can be controlled experimentally, including thermal annealing to cause crystallization or adding additives to modify the domain size of donor and acceptor. PTB7 and PC 70 BM are commonly used donor and acceptor for application in solution-processed OSCs, with repeatable cell performance and high PCE. The PTB7:PC 70 BM blend system was used in this study. Chlorobenzene (CB) solvent was used to dissolve the donor and acceptor materials. The additive Diiodooctane (DIO) was also used to control the morphology of active layer. Fig. 2.3 Chemical structures of PTB7, PC 70 BM, CB and DIO. 14

30 2.3 Characterization of OSCs Equivalent circuitry diagram It is convenient to describe the electrical behavior of a solar cell with an equivalent electric circuitry model. Fig. 2.4 shows the equivalent circuitry diagram of a solar cell. This model was originally constructed for inorganic solar cells, as an approximation, this model is also adopted for OSCs. Fig. 2.4 Equivalent circuitry diagram of a solar cell. In this diagram, the current source G represents the photocurrent generated by active layer in OSC. I G is the current generated by free charge carriers after they are separated at the D/A interface, but before they transport in the bulk. The diode D represents the asymmetric conductivity of an OSC, arising from the built-in potential in the bulk of an OSC. After the free charge carriers are generated, loss will occur before they are collected at the electrodes. Shunt resistance (R Sh ) and series resistance (R S ) are considered mainly from two different loss mechanisms. Rsh is related to the exciton recombination loss at the D/A interface. Therefore, I Sh is the current lost due to recombination at the D/A interface. Ideally, R Sh should be infinite, where no recombination occurs. However, excitons loss due to the traps, inefficient exciton 15

31 dissociation and recombination at the D/A interface contribute to a smaller R Sh. R S is related to charge carriers lost along the path of transportation and collection. For an ideal OSC, this value should be zero. In reality, charge carrier mobility in the bulk, energy barriers at electrode interfaces will result in an increase in R S. R L is the external load in the circuitry, which can also be a source meter used in current density voltage (J V) characteristics measurement Current density Voltage characteristics J V characteristics of an OSC are shown in Fig The dotted line represents the current density measured in dark and the solid line is the current density measured under illumination. The organic semiconductor materials have no intrinsic carriers, therefore, under dark condition, no current is observed unless the device is under forward bias. Photocurrent can be generated when the device is under illumination. If there is no external bias, the device is in short circuit condition. The maximum current density measured under this condition is called the short circuit current density (J SC ). This is the maximum output current the device can achieve. As the forward bias increases, the measured current will decrease. When the current becomes zero, the device is under open circuit condition. The bias voltage at this condition is called the open circuit voltage (V OC ). This is the maximum output voltage the device can achieve. When the bias voltage is between zero and V OC, there is a point where the output power density researches its maximum P Max. The bias voltage at this maximum power point is V Mpp and the corresponding current density is J Mpp. 16

32 Current density (ma/cm 2 ) J SC J MPP P Max = V MPP * J MPP Photocurrent Darkcurrent Voltage (V) 2.5 J V characteristics of an OSC. V MPP The power conversion efficiency of the OSC is the ratio of the maximum power output of the OSC P Max and the power density of solar irradiation P in, as shown in Equation 2.1. V OC Fig. Fill factor (FF) is another important factor to characterize solar cells. It is defined as Therefore, PCE can also be expressed as From equation 2.3 we can clearly see, since the power density of illumination P in is 17

33 fixed, PCE is directly determined by FF, V OC and J SC. To accurately measure the PCE of OSCs under solar irradiation in a laboratory, solar simulator is used. In reality, the spectrum and intensity of solar irradiation are different at different time during the day. The incident angle changes as the sun s position changes. Consequently, the spectrum varies as solar irradiation passes through a different thickness of earth s atmosphere. Light is absorbed and scattered by a different amount. Other variations in natural conditions can also affect the solar irradiation. A standard solar spectrum and intensity called AM1.5G solar irradiation of 100 mw/cm 2 is set by the photovoltaic industry for the purpose to measure the PCE of OSCs in the laboratory. AM stands for air mass coefficient. It is defined as the ratio between the direct optical path length as solar irradiation passes through the Earth's atmosphere and the path length when the sun is at the zenith, as expressed in equation 2.4. AM 1.5 corresponds to solar irradiation when θ equals to

34 Fig. 2.6 Definition of air mass. There are two distinct spectra: the AM 1.5 direct and AM 1.5 global (as shown in Fig. 2.7). AM 1.5 direct only considers the spectrum measured in the direction of the sun. AM 1.5 global takes into account the scattered light from other directions. AM 1.5 global is normally used for the characterization of solar cells because it is more relevant to the actual operating condition of solar cells. Fig. 2.7 shows the AM 1.5 Global spectrum and AM 1.5 Direct spectrum. 19

35 Fig. 2.7 Spectra of AM 1.5 Global and AM 1.5 Direct. A schematic diagram of the J V measurement system is shown in Fig Solar simulator acts as light source. A xenon lamp with an AM 1.5 filter is used to generate an AM 1.5G solar spectrum. Light intensity is calibrated using a NREL certified Si solar cell to 100 mw/cm 2. The simulated solar irradiation is perpendicular to the OSC substrate. A source meter is connected to the OSC to provide bias voltage. Finally, a computer is used to record the J V curve. 20

36 Fig. 2.8 Schematic diagram of J V measurement system setup IPCE measurement The incident photon-to-electron conversion efficiency (IPCE) is another important measurement for OSCs. It is also called external quantum efficiency (EQE). This measures the collected electron per incident photon at different wavelengths. The expression is given in equation 2.5. where h is the Planck s constant, c is the speed of light, P is the power of incident light, λ is the wavelength of the incident photon and e is the electron charge. The schematic diagram of the IPCE measurement system is shown in Fig A Xenon lamp is used as light source. The monochromator is used to generate monochromatic light. An optical chopper is connected to a lock-in amplifier and set to 21

37 a specific frequency. This is to amplify the current signal generated by solar cell resulting from the incidence of monochromatic light. Since lighting system in the laboratory uses an AC electrical power with frequency of 50 Hz, the frequency of the optical chopper should be set to a value that is not the frequency of 50 Hz to the avoid interference from room lighting. Fig. 2.9 Schematic diagram of IPCE measurement system setup. The power of the incident light P is not a constant since light sources in different labs are different, and even the same light source may give different intensity depending on external environment factors and warming up time. To avoid this variable, a calibrated Si standard solar cell with known IPCE spectrum is first measured. The power of incident light can be expressed as: Substitute equation 2.6 into 2.5, Now the only variables are I SC and I SC(Si), and they can be directly measured. 22

38 Chapter 3: Experimental 3.1 Materials process and characterization Optical properties of functional layers The absorption spectrum of the active layer was measured using the UV-Vis spectroscopy over the wavelength range from 300 nm to 900 nm, which is the response region of the PTB7:PC 70 BM system. The setup is shown in Fig Fig. 3.1 (a) Schematic diagram of a UV-Vis spectroscopy and (b) a photo picture of the UV-Vis spectroscopy used in this work. Fig. 3.2 shows the absorption spectra of PTB7:PC 70 BM blends with different weight ratios of PC 70 BM to PTB7. The measurements were performed using PTB7:PC 70 BM films spin-coated on glass substrates. The absorption of the active layers was obtained by subtracting the absorption of the active layer/glass sample by the absorption of the glass. The absorption spectra are normalized at the absorption peak of PTB7 at the wavelength of 680 nm. 23

39 Fig. 3.2 Absorption spectra of PTB7:PC 70 BM films with different weight ratios of PTB7:PC 70 BM in the blend. Other important optical parameters are refractive index n(λ) and extinction coefficient k(λ) of the functional materials used in the OSCs. They are used to calculate the optical properties of a layer structured device in the simulation. Details about the theoretical simulation will be discussed in chapter 4. Variable angle spectroscopic ellipsometry (VASE) was used to measure the wavelength dependent refractive index n(λ) and extinction coefficient k(λ) of the samples. The configuration of a VASE setup is shown in Fig Linearly polarized light is reflected from the sample. The amplitude ratio Ψ, and the phase difference Δ of the reflected light are measured. They correspond to the optical properties of the films. These two parameters along with the film thickness are fitted to determine the n(λ) and k(λ) using different models. Fig. 3.4 shows the n( ) and k( ) measured for the functional materials used in this thesis. They are obtained from VASE measurement for films coated on silicon substrates. 24

40 Fig. 3.3 Configuration of a VASE. Fig. 3.4 Wavelength-dependent n( ) and k( ) measured for functional materials e.g. ITO, blend layer of PTB7:PC 70 BM, ZnO and MoO x used in this work. 25

41 3.2.2 Sheet resistance and resistivity of AZO The sheet resistance of TCO is an important parameter for transparent electrode used in OSCs. Lower sheet resistance means better conductivity. Therefore, OSCs made with electrode having lower sheet resistance will have a better performance. The sheet resistance of TCO was measured with a four-point-probe. Fig. 3.5 shows the measured sheet resistance (black dots) of AZO films as a function of film thickness. Sheet resistance is linearly related to film thickness, thicker film will have a lower sheet resistance. As can be seen the sheet resistance decreases linearly with the film thickness. Another parameter is the resistivity. It is the product of sheet resistance and film thickness. Ideally, resistivity does not depend on the thickness of the film, but rather on the property of the material itself. The calculated values of resistivity of AZO films (red dots) as a function of the film thickness are plotted in Fig It can be seen that the resistivity of different AZO films remains the same over the film thickness range from ~800 nm to 900 nm, with film resistivity of ~ cm, which is about an order of magnitude higher than that of the ITO commonly used for OSCs. To compensate for the high resistivity, a thicker AZO (~850 nm) was used. This results in a sheet resistance of around14 Ω/sq for the AZO electrode, comparable to the sheet resistance of ITO electrode used in this work (~ 10 Ω/sq). 26

42 Sheet resistance (Ohm/sq) Resistivity (10-4 Ohm cm) AZO thickness (nm) Fig. 3.5 Sheet resistance & resistivity of AZO films as a function of film thickness. 27

43 Chapter 4: Optical and optimal design of organic solar cells 4.1 Optical admittance analysis Optical modeling is a powerful approach to analyze optical effects and optimize absorption in organic solar cells. An OSC can be considered as a multilayer thin film system consisting of a stack of organic and inorganic layers having film thicknesses in an order of a few hundred nanometers. Light is electromagnetic wave travelling inside the OSCs. As light travels through the OSCs, reflection and transmission takes place at the interfaces of two different layers due to the difference in the refractive index of the adjacent materials. Absorption will occur as light travels through an absorbing medium due to the extinction coefficient k(λ). The reflected light and the incoming light will cause interference, producing light distribution in the OSCs. Optical admittance analysis can be applied to calculate the distribution of light inside the OSCs [28,29]. If the thickness, the wavelength-dependent refractive index n( ) and extinction coefficient k(λ) of each layer are known, absorption, transmission and reflection of each layer can be calculated, therefore, light distribution in the OSCs can be obtained. The goal is to design the structure of OSCs in such a way that light distribution has its maximum intensity located in the active layer of the OSCs for maximum light absorption. Consider an OSC with m layers. The effective optical admittance y eff is defined as y eff = C/B where C and B are given as [30,31]: y j and y m+1 represents the admittance of the jth layer and the substrates respectively. I 28

44 is the unit matrix and δ j is the angular phase given by: N j is the complex refractive index of the jth layer, d j is the thickness of the jth layer. The total reflection of this m layer OSC is given by: where N 0 is the refractive index of air. At normal incidence the total transmittance is: where The absorption of the jth layer in this OSC can be expressed as: This means that by knowing the thickness, refractive index and extinction coefficient of each layer, we can calculate the wavelength-dependent reflection, transmission and absorption spectra within each layer of the OSC. Under the solar irradiance with flux F(λ), we can then calculate the integrated absorptance of any given layer under the sun. The integrated absorptance can be expressed as: 29

45 4.2 Device design and optical modeling The cross sectional views of conventional and inverted structured OSCs based on PTB7:PC 70 BM blend system are shown schematically in Fig The regular OSCs have a structure of ITO/PEDOT:PSS(40 nm)/ptb7:pc 70 BM/ZnO(10 nm)/ag(100 nm). Where ITO functions as an anode, PEDOT:PSS as a HTL and ZnO as a ETL. The inverted OSCs have a typical structure of ITO/ZnO(10 nm)/ PTB7:PC 70 BM/MoO x (2.0 nm)/ag (100 nm), where ITO acts as a cathode, ZnO as a ETL and MoO x as a HTL. Fig. 4.1 Cross sectional views of (a) conventional and (b) inverted OSCs based on PTB7:PC 70 BM blend system. Optical properties of both types OSCs were studied using optical admittance analysis. The thickness of the PTB7:PC 70 BM active layer was varied from 40 nm to 300 nm with an increase step of 20 nm. The results of integrated absorptance of the PTB7:PC 70 BM active layer as a function of its layer thickness, calculated with equation 4.8 are plotted in Fig The thickness of active layer has a dominant 30

46 Integrated absorptance (%) impact on its integrated absorptance, which can be adjusted by the process conditions. Thicknesses of interfacial layers such as ZnO, MoO x and PEDOT:PSS on the other hand, were kept constant of 10 nm, 2 nm and 30 nm, optimized experimentally for OSC fabrication. The thickness of rear reflecting Ag electrode was kept at 100 nm to ensure high reflectivity and good conductivity. In this work, the thickness of active layer used in the OSCs was changed in simulation Inverted Conventional Active layer thickness (nm) Fig.4.2 Calculated integrated absorptance of active layer as a function of its thickness for conventional and inverted OSCs. The results of integrated absorptance of active layer as a function of its layer thickness of both conventional and inverted structured OSCs, based on the PTB7:PC 70 BM blend system are shown in Fig.4.2. It can be seen that the integrated absorptance of the layer exhibits an oscillation behavior with increase in its layer thickness. Two relative absorption maxima occur at the blend layer thicknesses of 100 nm and 250 nm for 31

47 regular OSCs, and 110 nm and 270 nm for inverted OSCs. The local maxima correspond to relatively higher photocurrent generation in OSCs due to constructive light interference. To illustrate the effect of active layer thickness on integrated absorptance in OSCs, electric field within inverted OSCs with active layer thickness of 60 nm and 110 nm was calculated and plotted in Fig The wavelength of the incoming light is 500 nm. Fig. 4.3 Electric field distribution in PTB7:PC 70 BM-based inverted OSCs with active layer thickness of (a) 60 nm and (b) 110 nm. The distribution of electric field is produced by interference of reflected and incoming light. Since electric field strength in the active layer is proportional to light absorption, a stronger E field will result in higher absorption. It can be seen that the electric field 32

48 in the active layer is stronger for the OSCs with 110 nm active layer than the OSC with 60 nm active layer. It is known that light absorption in OSCs is limited due to the presence of a mismatch between optical absorption length and charge transport scale, caused by the low charge mobility in conjugated polymers. Therefore, in this case, PTB7:PC 70 BM thickness of 100 nm for conventional structure and 110 nm for inverted structure are the optimal thickness for device fabrication. It should be noted that the integrated absorptance of inverted OSC at its optimal active layer thickness of 110 nm is 70%, higher than that of the conventional structure at its optimal active layer thickness of 100 nm, which is 67%. Furthermore, for the majority part of the integrated absorptance curve, inverted structure demonstrates higher absorption in active layer than the conventional structure. This may be caused by the favorable light distribution in the inverted structure, and the absorption cost due to PEDOT:PSS in conventional structure. This result suggests that light harvesting in inverted structure is better that the conventional structure. Our experimental results agree with the theoretical simulation in showing that the inverted OSCs having a 110 nm thick PTB7:PC 70 BM layer correspond to the best cell performance. Moreover, the fabrication of conventional OSCs using AZO as the anode faces a technical difficult. The acidic nature of the PEDOT:PSS layer would induce a gradual deterioration in the properties of the acid-sensitive AZO, resulting in a poor device performance and device lifetime. Based on the above reasons, inverted AZO-based OSC was used for the study. 33

49 Chapter 5: Organic solar cells with AZO cathode 5.1 Recent progresses in AZO-based OSCs ZnO-based TCO is a promising candidate for ITO replacement in OSCs. When doped with group III elements such as Al or Ga, the conductivity of ZnO can be greatly enhanced. Therefore, Al-doped ZnO (AZO) and Ga-doped ZnO (GZO) have attracted a lot of attentions during the past few years. However, the performance of OSCs based AZO is less satisfactory with PCE of ~2%, much lower than that of the cells made with ITO anode. In conventional structure, AZO will be used as an anode. Since the work function of AZO is usually low compared to that of ITO, energy alignment at the organic/azo interface is not favorable for hole extraction. In order to increase the work function of AZO, HTL must be incorporated. PEDOT:PSS is the most commonly used HTL. But AZO is very sensitive to acid and the acidic PEDOT:PSS will deteriorate the AZO properties. Therefore, AZO based OSCs with inverted structure are preferred, in which AZO serves as a cathode. Therefore, its low work function is favorable for the energy alignment. Furthermore, avoiding the use of acidic PEDOT:PSS helps to improve the device stability. This is supported by the fact that the overall performance of inverted AZO-based OSCs is better than that of AZO-based regular OSCs. Table 1 summarizes the reported performance of AZO-based OSCs. The result of this thesis was published in the journal Applied Physics Letters and is shown in the shaded row in Table 1 [43]. 34

50 Table 1 Summary of results on OSCs using AZO as electrode Device Structure PCE Structure Ref. AZO/Au/ CuPc/C 60 /Alq 3 /Al 1.40% Conventional [32] AZO/CuPc/C 60 /TPBI/Al 1.30% Conventional [33] AZO/PEDOT:PSS/P3HT:PCBM/Ca/Al 2.01% Conventional [34] AZO/Ag/AZO/PEDOT:PSS/P3HT:PCBM/Ca/Al 2.14% Conventional [35] AZO/Ni/NiO/ P3HT:PCBM/LiCoO 2 /Al 2.36% Conventional [26] Ag/AZO/Ni/MoO 3 /ZnPc:C 60 /C 60 /Bphen/Ag 2.6% Conventional [36] AZO/Ag/AZO/TiOx/P3HT:PCBM/PEDOT:PSS/Ag 2.07% Inverted [37] ITO/AZO/P3HT:PCBM/NiOx/Ag 1.40% Inverted [38] AZO/Al//LiCoO 2 /PSBTBT:PC 70 BM/PEDOT:PSS//ITO 3.0% Inverted [39] AZO/P3HT:PCBM/MoOx/Ag 3.06% Inverted [40] ITO/AZO/N719/P3HT:PCBM/WO 3 /Al 3.80% Inverted [41] AZO/Al/TiOx/PCBTDT:PC70BM/ PEDOT/ITO 4.0% Inverted [42] AZO/ZnO/PTB7:PC 70 BM/MoO X /Ag 6.15% Inverted [43] AZO/Al/P3HT:ICBA/MoO X /Al 5.52% Inverted [44] AZO/ZnO/P3HT:PCBM/MoO X /Ag 3.01% Inverted [45] It can be seen that the overall PCE of AZO-based inverted OSCs, although is less than 4%, is higher than the conventional OSCs using AZO anode. This suggests that AZO is quite suitable for application in OSCs. In inverted OSCs, it was reported and widely accepted that the presence of an ETL between the front transparent cathode and the organic donor/acceptor blend enables to reduce the exciton recombination at the 35

51 interface and therefore further enhance the performance of the OSCs. Enhanced performance of inverted OSCs using ETL-modified ITO front cathode, e.g., a thin layer of solution-processed ZnO or TiO x has been demonstrated [46,47]. However, the report on the success in the development of high performance inverted OSCs using AZO front transparent cathode is quite limited, as compared to the significant progresses made in ITO-based OSCs. 5.2 Modification of AZO surface with solution-processed ZnO A set of OSCs with a structure of glass/azo/zno/ptb7:pc 70 BM/MoO x /Ag was fabricated. An ITO-based inverted OSC having an identical structure of glass/ito/zno/ptb7: PC 70 BM/MoO x /Ag was used as a control cell for comparison studies. An 850 nm thick AZO film on glass (Xinyi Glass) has a sheet resistance of 14 /square and a 170 nm thick ITO having a sheet resistance of 10 /square were used for OSC fabrication. The AZO/glass and ITO/glass substrates were cleaned in ultrasonic bath with detergent and subsequently rinsed in ultrasonic bath with deionized water, acetone and isopropanol in sequence. ZnO nanoparticles with particle diameter around 5.0 nm were dissolved in methanol. The ZnO nanoparticles were synthesized following the processes described in a previous work [48]. To avoid possible aggregation of ZnO nanoparticles, the solution was subjected to a 15 min ultrasonic bath before use. A layer of 10 nm thick ZnO ETL was then coated on ITO/glass and AZO/glass by spin-coating inside a N 2 -purged glove-box with O 2 and H 2 O levels < 0.1 ppm. The PTB7 (donor):pc 70 BM (acceptor) blend in a weight ratio 36

52 of 1:1.3 was dissolved in chlorobenzene (CB) (Sigma-Aldrich 99.8%) with 3% 1, 8-Diiodooctane (DIO) (Sigma-Aldrich). PTB7 polymer was purchased from 1 Material and PC 70 BM from Nano C. All chemicals were used as received. The solution of donor/acceptor blend was stirred on a hotplate at 70 0 C before use. A 100 nm thick polymer blend layer was then deposited on ZnO-modified AZO/glass and ITO/glass substrates by spin-coating inside the glove-box. The samples were then transferred to a connected vacuum system, with a base pressure of Pa, for deposition of a 2.0 nm thick MoO x anode interlayer and a 100 nm thick Ag top contact without exposing them in air. The OSCs were then transferred back to the glove-box for J V characteristic measurement under a SAN-EI XEC-301S AM 1.5G solar simulator with calibrated power density of 100 mw/cm 2. The measuring source unit used was Agilent U2722 SMU. The thicknesses of ZnO ETL and polymer blend layer were measured with a Veeco Dektak 150 surface profiler. MoO x and Ag thicknesses were monitored in-situ using a calibrated Fil-Tech QI8010 quartz crystal microbalance. For the UV-assisted acceleration aging test, a mercury vapor bulb (LOCTILE UVALOC 1000 UV system housed inside the glove box) with an intensity of about 90 mw/cm 2 was used. The glass substrate is 2.5 cm 2.5 cm in size. Both ITO and AZO share the same pattern design as shown in Fig

53 Fig. 5.1 The pattern of ITO/AZO on glass substrate. After forming the active layer and other buffer layers, shadow mask was used for forming the top by thermal evaporation, as shown in Fig Fig. 5.2 The design of a completed organic solar cell. The blue area represents the rear electrode. Shaded area represents the active cell area, which is the overlapping area between the rear electrode and the under-lying ITO/AZO contact. There are four identical individual cells on one substrate as illustrated. The ITO/AZO pattern and shadow mask were designed in such a way that each cell has the same area of 3.0 mm 3.0 mm = 9.0 mm 2. The plus and minus signs indicate the anode and cathode connection respectively in a regular OSC. Their 38

54 functions will be switched in an inverted OSC. J V characteristics of four OSCs were measured simultaneously to avoid the differential aging problem encountered typically in the sequential measurement for each cell. The cross-sectional view of multilayer inverted OSCs of the type: glass/front cathode/ptb7:pc 70 BM (110 nm)/moo x (2.0 nm)/ag (100 nm) is shown in Fig.5.3. The thickness of the active layer was optimized to be 110 nm based on experimental results and optical optimal modeling data discussed in Chapter 4. The thickness of MoO x and ZnO was also optimized experimentally to be 2.0 nm and 10 nm respectively. Three different front cathodes of AZO/ZnO(10 nm), ITO/ZnO(10 nm) and bare AZO, as illustrated in Fig.5.3, were used for fabrication of the inverted OSCs. Fig. 5.3 Cross-sectional view of an inverted OSC made with a front transparent cathode. 39

55 5.3 Device characterization and analysis The J V characteristics measured for a set of structurally identical inverted OSCs made with three different cathodes of AZO/ZnO (10 nm), ITO/ZnO (10 nm) and bare AZO are plotted in Fig The parameters of the inverted OSCs with different cathode contacts are summarized in Table 2. It can be seen that OSCs made with an AZO/ZnO (10 nm) cathode has a PCE of 6.15%, which is slightly lower than 6.57% of a control OSC made with an ITO/ZnO (10 nm) cathode. OSCs with 10 nm thick ZnO-modified AZO and ITO cathodes have the same V OC of 0.74 V and FF of 66%. It is clear that a bare AZO cathode is not an efficient cathode contact for inverted OSC. As the OSCs have the same device configuration, the identical V OC and FF measured for OSCs with different cathodes of AZO/ZnO (10 nm) and ITO/ZnO (10 nm) reveal that both cells have similar electron extraction properties, determining mainly by the 10 nm thick ZnO interlayer. ITO and AZO layers act primary as the transparent electrode for lateral conductivity, implying the suitability of AZO as an ITO replacement for application in OSCs. 40

56 Fig. 5.4 J V characteristics of inverted OSCs made with different cathodes of AZO (850 nm)/zno (10 nm), ITO (170 nm)/zno (10 nm) and bare AZO (850 nm), measured under simulated AM1.5G, 100 mw/cm 2 illumination. Table 2 Summary of the performance of OSCs made with different transparent cathodes of AZO/ZnO, ITO/ZnO and AZO Front cathode V OC (V) J SC (ma/cm 2 ) FF (%) PCE (%) AZO/ZnO ITO/ZnO AZO The wavelength dependent transmittance, T( ), of AZO/glass and ITO/glass used in this work were measured and are plotted in Fig It shows that AZO/glass has an average transmittance of 79% over the wavelength range from 380 nm to 780 nm, which is slightly lower than 84% of ITO/glass. The fringes seen in the T( ) of AZO/glass is due to light interference effect caused by a thick AZO film. As AZO has a relatively poor bulk conductivity as compared to ITO, therefore an ~850 nm thick 41

57 AZO layer was chosen based on a consideration of achieving high optical transparency of ~80% in the visible light wavelength and a low sheet resistance of ~14 /square for OSCs. It can also be seen obviously that T( ) of AZO exhibits a cutoff at a short wavelength around 380 nm. This is caused by UV absorption of ZnO (band gap ~3.0 ev), which is the main component of AZO. The cutoff at around 380 nm in T( ) of AZO explains well the reduced J SC of 12.4 ma/cm 2 measured for an AZO-based OSC as compared to 13.4 ma/cm 2 of an ITO-based control OSC. As the plot of AM1.5G solar spectrum in Fig. 5.5, the photons with wavelength below 380 nm are able to enter into an ITO-based control OSC contributing to the photocurrent generation, while AZO front cathode in an AZO-based OSC behaves like a UV filter, blocking short wavelength light penetrating into the OSCs. Fig. 5.6 shows the IPCE of OSCs made with AZO/ZnO (10 nm) and ITO/ZnO (10 nm) front transparent cathodes. It can be seen that IPCE spectrum of the AZO-based OSCs has a red shift in the short wavelength side, which corresponds to the cutoff in T( ) of AZO/glass as shown in Fig Although a control OSC with an ITO front cathode has relativity high spectral response at the short wavelength region, the main energy of photons in the visible part of solar irradiation falls in the spectrum with wavelength > 380 nm. Apart from the UV-portion of the solar irradiation that is blocked by the AZO front electrode, it allows most of sunlight to pass through, enabling the efficient absorption in the active layer in the OSCs as ITO does. The fringes observed in the measured IPCE spectrum of AZO-based OSCs also consist with the T( ) measured for the AZO/glass, as shown in Fig

58 Fig. 5.5 Transparency measured for an AZO/glass and an ITO/glass over the wavelength range from 320 nm to850 nm, spectrum of AM1.5G solar irradiance is also plotted. Fig. 5.6 IPCE of inverted OSCs made with an AZO/ZnO (10 nm) transparent cathode, and a control OSC having an ITO/ZnO (10 nm) cathode. 43

59 Fig. 5.7 Schematic energy diagram of an AZO-based inverted OSC. The schematic energy diagram of an AZO-based inverted OSC is illustrated in Fig It shows that ZnO has a deep valence band level, serving as an effective holeblocking layer and acting as an electron extraction layer, allows effectively extracting electrons and suppress electron-hole recombination, leading to higher V OC and FF [46]. It shows that OSCs with ZnO-modified AZO front cathode have a PCE of 6.15% as compared to 4.28% measured for a structurally identical inverted OSC on a bare AZO contact, shown in Fig The enhancement in PCE of the cells mainly comes from the higher V OC and FF. This demonstrates that ZnO ETL plays a critical role in determining the performance of AZO-based inverted OSCs due to its efficient hole blocking effect, resulting in a reduced exciton recombination at cathode/blend layer interface. In OSCs, the photo-excited electrons and holes, generated from the dissociation of photo-generated excitons, are driven towards the corresponding 44

60 cathode and anode under the built-in potential. However, due to the structural configuration of donor/acceptor blend in a BHJ type OSC, both p-type polymer and n-type fullerene derivative are in contact with the cathode. If holes are effectively blocked at cathode/blend layer interface, the interfacial recombination will less likely to occur. To better understand the performance enhancement in an inverted OSC due to ZnO ETL layer, morphological properties of AZO/ZnO and AZO substrates were also investigated. Figs. 5.8 (a) and (b) are the atomic force microscopy (AFM) images measured for a bare AZO and a 10 nm thick solution-processed ZnO-modified AZO substrates. Fig. 5.8 AFM images measured for (a) glass/azo and (b) glass/azo/zno. Results show that a bare AZO film has a RMS roughness of 14.3 nm and that for a 10 nm thick ZnO-modified AZO surface has a RMS of 13.1 nm. They have similar surface morphology with root mean square (RMS) roughness of 14.3 nm and 13.1 nm respectively. The ZnO nanoparticles may partially fill the valleys between grains of the AZO film, resulting in a slight reduction in RMS roughness. But the difference of roughness is very small, which implies that the slight 45

61 change observed in the surface RMS is not the primary reason responsible for better performed inverted OSCs with an AZO/ZnO (10 nm) cathode. Therefore, the enhancement in the cell performance is mainly due to the improved electronic properties, such as hole-blocking ability by the ZnO interlayer as discussed above. 46

62 5.4 Stability of AZO-based organic solar cells Apart from efficiency and cost competitiveness, stability is another important factor with a practical implication if OSC is to become a viable option for sustainable green energy source. It has been reported that UV exposure would cause a continuous deterioration in the performance of OSCs due to an inevitable UV-induced degradation process in the functional active layer [49]. Although in real application, moisture and oxygen are also possible sources of OSC degradation, these influences can be minimized by incorporating a proper encapsulation. But the UV-induced degradation can only be eliminated by preventing UV light in solar irradiance reaching the active region in OSCs. This involves the use of a UV filter in OSCs, which complicates manufacturing process and also adds additional cost. From T( ) of an AZO film (shown in Fig. 5.5) it can be seen evidently that AZO has a poor transparency in UV range, indicating the possibility of using it as a natural UV filter, which can be beneficial to prolong the OSC lifetime. In order to manifest this advantage, UV-assisted acceleration aging test was carried out. A pair of structurally identical OSCs made with two different cathodes of AZO/ZnO and ITO/ZnO were encapsulated and exposed to UV irradiation for different exposure time from 0 20 min through the glass side. The UV-assisted acceleration aging test was conducted inside the N 2 -purged glove-box with O 2 and H 2 O levels < 0.1 ppm, therefore the possible OSC degradation due to the encroachment of moisture and oxygen can be neglected. The variation in cell parameters including PCE, FF, J SC and V OC measured for the inverted OSCs with AZO and ITO (control) under different UV exposure times 47

63 was recorded, the measured cell parameters are plotted in Fig Fig. 5.9 Recorded PCE, FF, J SC and V OC of inverted OSCs made with ITO and AZO front cathodes versus the UV exposure time. It can be seen that although a control ITO-based OSC had a slightly higher starting PCE, a fast decrease in all cell parameters of PCE, FF, J SC and V OC was observed under the UV exposure. After a 20 min continuous UV exposure, V OC, FF, and J SC of the control OSC dropped from 0.74 V, 67% and 13.2 ma/cm 2 to 0.58 V, 49% and 11.5 ma/cm 2, resulting in a PCE falling from 6.53% to 3.3%. For an AZO-based OSC, however, it only experienced a moderate decrease in the cell performance under the same UV-assisted acceleration aging test. The results reveal that V OC changed from 0.75 V to 0.72 V, FF decreased slightly from 63% to 59%, while J SC remained almost unchanged, from 12.4 ma/cm 2 to 12.3 ma/cm 2, resulting in an overall PCE of AZO-based OSCs only had a minor decrease from 5.86% to 5.2% after a 20 min continuous UV exposure. The distinctive enhancement in the stability of AZO-based 48

64 OSCs arises from the tailored absorption of AZO electrode in wavelength < 380 nm, serving as a UV filter to inhibit an inevitable degradation in ITO-based OSCs caused by the UV exposure. In addition to the optical and electric properties, our results demonstrate clearly that AZO film can be a very good ITO alternative for application in OSCs where exposure to UV light is unavoidable. 5.5 Charge extraction efficiency of AZO-based OSCs The results of this work indicate that AZO is a suitable electrode material for high performance OSCs. The use of AZO/ZnO allows improving FF in inverted OSCs. This suggests that AZO/ZnO cathode in reverse configuration OSCs has good charge extraction property, comparable to OSCs made with ITO/ZnO cathode. To verify that and further investigate the charge extraction properties of AZO-based OSCs, photocurrent density (J ph ) effective voltage (V eff ) characteristics and light intensity dependent J ph V eff for both AZO-based and ITO-based inverted OSCs were examined and analyzed. Photocurrent density J ph is defined as: where J L is the current density generated under illumination and J D is the current density measured in dark. Effective voltage V eff is defined as: where V b is the bias voltage and V 0 is the compensation characterized as the voltage when photocurrent J ph equals to 0. Therefore, V 0 is usually slightly larger than V OC. 49

65 Current density (ma/cm 2 ) Fig illustrates the definition of V J SC 0 0 V OC V Photocurrent Darkcurrent -40 J ph = J L - J D = Voltage (V) Fig Illustration of compensation voltage V 0. J ph V eff characteristics of inverted OSCs made with AZO/ZnO cathode and ITO/ZnO cathode are shown in Fig Ideally, J ph should be independent of external bias voltage, represented by the dotted lines in Fig. 5.11, give by [50]: where e is the electron charge, G is the generation rate of charge carriers and L is the thickness of active layer. However, when the effective voltage across the device is low, recombination is more likely to occur. This will reduce the photocurrent density. It can be seen from Fig 5.11 that J ph gradually increases with V eff and saturates at V eff >1 V. The saturated J ph of OSC made with ITO/ZnO cathode was slightly larger than that of the OSC made with AZO/ZnO cathode. This agrees with the previous results. At low effective field (0.01 V < V 0 V b < 0.1 V) however, J ph of OSCs made with AZO/ZnO 50

66 J ph (ma/cm 2 ) cathode is comparable and even slightly larger than that of OSCs made with ITO/ZnO cathode. This result suggests that even at low electric field, charge extraction for AZO-based OSCs is still comparable to ITO-based OSCs. 0.1 AZO/ZnO ITO/ZnO V 0 -V b (V) Fig J ph V eff characteristics of inverted OSCs made with different cathodes of AZO/ZnO and ITO/ZnO. Light intensity dependent photocurrent density was also measured. It was reported that J ph has a power dependence on light intensity I, i.e., J ph I a [51]. For an ideal OSC with efficient charge extraction and no space charge effect, J ph is linear to the light intensity and a = 1. However, charge carrier loss due to bimolecular recombination causing the factor a derivates from 1. Due to unbalance charge mobility in organic materials, carriers with low mobility (e.g., holes) will accumulate in the device and cause space charge limit effect. When J ph is space charge limited, the following relation holds [50]: 51

67 J ph (ma/cm 2 ) where μ h is the mobility of holes in the active layer. Since the charge generation rate G is directly related to light intensity I, the space charge limited photocurrent results in the factor a of Therefore, most OSCs has a ranging between 0.75 and 1. J ph V eff characteristics of inverted OSCs made with cathodes of AZO/ZnO and ITO/ZnO were measured under solar simulator with light intensity set to 23 mw/cm 2, 37 mw/cm 2, 58 mw/cm 2 and 90 mw/cm 2 and shown in Fig The plot of J ph as a function of light intensity is shown in Fig AZO/ZnO 23mW/cm 2 AZO/ZnO 37mW/cm 2 AZO/ZnO 58mW/cm 2 AZO/ZnO 90mW/cm 2 ITO/ZnO 23mW/cm 2 ITO/ZnO 37mW/cm 2 ITO/ZnO 58mW/cm 2 ITO/ZnO 90mW/cm V 0 -V b (V) Fig J ph V eff characteristics of inverted OSCs made with different cathodes of AZO/ZnO and ITO/ZnO. 52

68 Fig J ph as a function of light intensity for OSCs made with cathodes of (a) AZO/ZnO and (b) ITO/ZnO. The slopes of fitted lines and corresponding V eff are also shown. It can be seen from Fig that the slope, corresponding to the value of a obtained for both AZO-based and ITO-based OSCs are similar. At V eff = 1.80 V, a of both OSCs are 1, which means no bimolecular recombination occurs and all dissociated charges are extracted and collected. At very weak effective voltage of V eff = 0.10 V, a of AZO-based OSC is 0.88 and ITO-based OSC is This indicates that very little space charge is induced, and the charge extraction efficiencies for both OSCs are quite efficient. This result proves that with the modification of ZnO, AZO/ZnO cathode acts likely as an ITO/ZnO cathode with the same charge extraction property. The only difference lays in the poor transmittance of AZO in the wavelength < 380 nm, which slightly reduces the photocurrent generation in the active layer, but beneficial for enhancement in the lifetime of OSCs. 53

69 5.6 Correlation between the change in charge collection characteristics and the stability of the OSCs J ph V eff characteristics and light intensity dependent photocurrent of AZO-based and ITO-based inverted OSCs were also measure after the OSCs were exposed to UV irradiation from the cathode side in the glove-box for 10 mins. The recorded J ph V eff characteristics are shown in Fig Light intensity dependent photocurrent density is shown in Fig Fig Recorded J ph V eff characteristics of AZO-based and ITO-based OSCs under different light intensity (a) before UV exposure and (b) after 10mins UV exposure. Compare Fig (a) and (b), it can be seen clearly that before UV exposure, both types of OSCs had photocurrent approaching saturation at Veff of ~0.3 V. After 10 mins UV exposure, photocurrent of OSC made with ITO/ZnO cathode rapidly decreased in the range of Veff between 0.1 V and 1 V, while it remained unchanged for the OSC made with AZO/ZnO cathode. 54

70 Fig J ph as a function of light intensity for OSCs made with cathode of (a) AZO/ZnO before UV exposure, (b) AZO/ZnO after 10 mins UV exposure, (c) ITO/ZnO before UV exposure and (d) ITO/ZnO after 10 mins UV exposure. The slopes of fitted lines and corresponding V eff are also shown. As it is also shown in Fig. 5.15, slope of the factor a for AZO-based and ITO-based OSCs were similar before UV exposure. However, after 10 mins UV exposure, the factor a for ITO-based OSC dropped more quickly as compared to the AZO-based OSCs. This indicates that one of the reasons for the degradation of OSCs after being exposed to UV light is the increased chance of recombination in the OSCs. This may be caused by the UV-induced traps in the bulk of organic materials or at the interfaces which provide sites for exciton recombination and reduces the photocurrent. Since AZO has a UV filtering effect, UV cause degradation on the functional layers in OSCs made with AZO cathode is minimized, making AZO-based OSCs more stable under UV illumination. 55